mystery sequence (easy?)

[Max Alekseyev]

I've got a reply from the author, and that is how he explained his sequence:

Let T(n,k) be the n-th k-almost prime. Then
A109636(n) = T(n,k) such that k is minimum and for all l>0, T(n,k+l) >
2^l * T(n,k).

I did not check if that makes sense. But the author also noted that
the sequence now looks senseless to him, and probably should be
removed from OEIS.

Max

On 3/15/07, N. J. A. Sloane <njas at research.att.com> wrote:
> Dear Seqfans:
>
> Can anyone figure this out?
>
> %I A109636
> %S A109636 2,3,9,10,27,28,30,81,84,88,90,100,104,243,252,264,270,272,280,300,304,
> %T A109636 312,729,736,756,784,792,810,816,840,880,900,912,928,936,992,1000,1040,
> %U A109636 2187,2208,2268,2352,2368,2376
> %N A109636 Minimal numbers in k-almost prime numbers massiv of (n,k) such that number of (n,k+l) is equal 2^l times the number of (n,k) for any l>0.
> %H A109636 Wikipedia, <a href="http://en.wikipedia.org/wiki/Almost_prime">k-almost prime numbers</a>.
> %Y A109636 Cf. A000040, A001358, A014612, A014613, A014614.
> %Y A109636 Sequence in context: A047358 A003140 A057234 this_sequence A057236 A063257 A103039
> %Y A109636 Adjacent sequences: A109633 A109634 A109635 this_sequence A109637 A109638 A109639
> %K A109636 nonn,uned,obsc
> %O A109636 1,1
> %A A109636 Yury V. Shlapak (shlapak(AT)imp.kiev.ua), Aug 04 2005
>
> Possibly one of the Russian-speaking members of the list can
> help by translating it to Russian, where it might be clearer?
> What does "(n,k)" mean???
>
> Neil
>
> PS The "binary splitting" mystery sequence A126119 of the other day remains
> a mystery
>